hyungi_document_server/evals/markdown/pilot_1d_eval.csv at b09687d41d2761e707b2452a51fcf3e571f8fcbd

Files

T

Hyungi Ahn b09687d41d feat(scripts): Phase 1D Round 2 — controlled backfill stratification

기존 phase1d_pilot.py (단순 ai_domain × file_size 3-bucket) 를 plan
~/.claude/plans/stratified-mingling-otter.md 의 4축 + sample_source 분리
+ forced_include 로 augment.

Round 1 (ai_domain × file_size 3-bucket) 의 한계:
  pending PDFs 의 자연 분포만 반영 → 알려진 약점 (필기/스캔/한중일
  mixed OCR) 이 sample 에 안 들어옴. 1C 시각 확인에서 doc 4809
  (Note_240805_용접교육 필기) 가 실제로 그 패턴을 보였는데, 자연
  selection 에 맡기면 다음 라운드도 같은 case 가 빠질 위험.

Round 2 디자인:
  - 4 축 stratification: doc_type × file_size_band × text_density_band
    × handwritten_hint
  - sample_source ∈ {existing_success(5), controlled_backfill(25)}
  - forced_include doc 4809 — known bad anchor. 다음 튜닝/대안 도입 후
    같은 문서 재변환 결과와 1:1 비교 가능.
  - text_density = LENGTH(extracted_text) / (file_size / 1024) chars/KB
    가장 깨끗한 단일 proxy. 0.17(필기 4809) ↔ 94(born-digital 3759)
    양 끝 검증.
  - script_mix proxy: Hangul/CJK/Hiragana/Katakana/Latin Unicode block
    ratio → korean_dominant / mixed_korean_cjk / mixed_korean_latin /
    cjk_dominant / latin_dominant / unknown.
  - page_count_estimate: existing_success 는 md_extraction_quality.
    metrics.source_page_count 사용. controlled_backfill 은 NULL
    (marker 가 PyMuPDF 로 어차피 다시 읽음).
  - 시드 SAMPLE_SEED=20260502 고정, 재현성 보장.

Sample 분포 (실측 2026-05-02):
  bucket_label: born_digital=12, mixed=5, existing_calibration=4,
                handwritten=3, scan_likely=3, large=2, existing_anchor=1
  doc_type: Academic_Paper=7, study_note=6, Standard=5, Note=4,
            Reference=3, Manual=3, Drawing=1, Report=1
  file_size_band: M=14, S=12, L=4
  text_density_band: born-digital=15, scan-likely=9, mixed=6
  handwritten_hint: lo=26, hi=4 (모집단 1.1% 대비 13배 over-sample)
  forced anchor doc 4809 = density 0.17 (사용자 시각 확인의 그 문서)

새 subcommand:
  eval_template — pilot_1d_eval.csv 스켈레톤 (rubric 5축 1~5 +
  overall_pass + notes). 사용자가 MarkdownDoc + PDF 토글 비교하며
  점수 채움.

기존 cmd_enqueue (snapshot/backup/dedup) + cmd_report (quality 메트릭)
는 유지.

산출물:
  scripts/phase1d_pilot.py — 4축 + sample_source + forced_include +
    eval_template subcommand. CSV+JSON dual output.
  evals/markdown/README.md — rubric + decision matrix + workflow guide.
  evals/markdown/pilot_1d_sample.csv — 30 rows × 15 cols (시드 결과,
    재현성 보존).
  evals/markdown/pilot_1d_eval.csv — 빈 스켈레톤 (사용자 평가 후 채움).

실행 경계:
  Step 1~3 (selection / template / dry-run) = 본 PR 으로 완료.
  Step 4 (--yes enqueue, 실제 30건 markdown 큐 인입) = 사용자 timing
  승인 + 야간 단발 sweep 윈도우 (23:00~03:00 KST) 안에서 별도 실행.
  marker-service BATCH_SIZE=1, 30건 평균 5분/건 ≈ 2.5h.

Verify:
  GPU 서버 fastapi 컨테이너에서 select 실행 → 30건 sample CSV 생성됨.
  eval_template subcommand 동작 확인. enqueue dry-run 으로 30 doc_ids
  + snapshot 출력 후 사용자 취소 분기 확인.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-02 16:15:09 +09:00

4.1 KiB

Raw Blame History

1	doc_id	title	sample_source	bucket_label
2	4809	Note_240805_용접교육 필기	existing_success	existing_anchor
3	5248	작업자 재난안전사고 예방을 위한 위험성평가 기법 연구	existing_success	existing_calibration
4	4068	공업역학 동역학(제13판)_Chapter 21 3차원 강체 운동역학	existing_success	existing_calibration
5	5189	VIII-1_08-UB	existing_success	existing_calibration
6	5141	Structural Analysiss and Design of Process Equipment_00_Contents	existing_success	existing_calibration
7	4815	Note_240830_소음진동교육 필기	controlled_backfill	handwritten
8	4798	Note_240528_다이아프람워크숍	controlled_backfill	handwritten
9	4813	Note_240827_필기	controlled_backfill	handwritten
10	5151	THE PIPE FABRICATORS BLUE BOOK	controlled_backfill	scan_likely
11	5268	황현필의 진보를 위한 역사_6장 제주4-3사건의 왜국을 멈추라	controlled_backfill	scan_likely
12	5127	표준기계설계(KS)_08_핀	controlled_backfill	scan_likely
13	8855	2월 26일	controlled_backfill	mixed
14	4061	공업역학 동역학(제13판)_Chapter 14 질점의 운동역학_일과 에너지	controlled_backfill	mixed
15	3782	Safety and Health for Engineers_02_5 Local, International, and Voluntary Laws, Regulations, and Standards	controlled_backfill	mixed
16	5179	Hydrogen-Embrittlement	controlled_backfill	mixed
17	5133	압력용기 핸드북_기타	controlled_backfill	mixed
18	3757	Industrial Safety and Health Management(7-ED)_2 Development of the safety and Health Function	controlled_backfill	born_digital
19	3758	Industrial Safety and Health Management(7-ED)_3 Concepts of Hazard Avoidance	controlled_backfill	born_digital
20	5163	국내 지속가능경영보고서의 노동인권 분야에 대한 실태 분석	controlled_backfill	born_digital
21	5167	우리나라 기업의 환경정보 공시 현황과 제도적 개선방안	controlled_backfill	born_digital
22	5154	국내 금속가공 중소기업의 스마트팩토리 활용 정도에 대한 실증적 연구	controlled_backfill	born_digital
23	5155	스마트 팩토리의 전략적 활용 연구	controlled_backfill	born_digital
24	5137	Pressure Vessel Design Manual_01 General Topics	controlled_backfill	born_digital
25	5211	PTB-4-2013_00_Foreword	controlled_backfill	born_digital
26	5178	Hydrogen_Piping_and_Pipelines_ASME_Code	controlled_backfill	born_digital
27	5168	TCoYourPaperlessOffice-4.0	controlled_backfill	born_digital
28	3765	Industrial Safety and Health Management(7-ED)_10 Environmental Control and Noise	controlled_backfill	born_digital
29	3769	Industrial Safety and Health Management(7-ED)_14 Materials Handling and Storage	controlled_backfill	born_digital
30	5274	황현필의 진보를 위한 역사_12장 대한민국의 정신을 훼손하지 말라	controlled_backfill	large
31	5180	ASME Sec I 2025	controlled_backfill	large

4.1 KiB Raw Blame History

4.1 KiB

Raw Blame History